CN114286375A

CN114286375A - Mobile communication network interference positioning method

Info

Publication number: CN114286375A
Application number: CN202111545764.5A
Authority: CN
Inventors: 王苑汀; 叶文; 程鹏; 熊胜相
Original assignee: Beijing University of Posts and Telecommunications
Current assignee: Beijing University of Posts and Telecommunications
Priority date: 2021-12-16
Filing date: 2021-12-16
Publication date: 2022-04-05
Anticipated expiration: 2041-12-16
Also published as: CN114286375B

Abstract

The invention discloses a mobile communication network interference positioning method, belonging to the field of positioning of an unknown interference source outside a network; firstly, selecting a road measuring point to perform reverse ray tracking, simulating signal propagation tracks from an interference source to each starting point, and screening similar tracks; then, dividing the search area of the interference source into three-dimensional grids, taking the grids as nodes, and forming an edge between two grids through which the track continuously passes to form a network model. And calculating the attribute value of each node according to the grid index to form an initial node set, filtering out nodes which cannot be an interference source according to a filtering rule, and leaving a candidate node set. And calculating the influence of each candidate node and the propagation probability of all edges corresponding to each node, carrying out influence propagation in the network model, and selecting the node with the maximum propagation influence as an interference source positioning position. The invention ensures that the more ideal positioning precision is achieved and simultaneously reduces the investment of manpower and material resources.

Description

Mobile communication network interference positioning method

Technical Field

The invention belongs to the field of positioning of an unknown interference source outside a network for analyzing network propagation influence, and particularly relates to a mobile communication network interference positioning method.

Background

In recent years, various mobile communication networks such as 4G, 5G, wireless private networks and the like have been developed rapidly, and great convenience is brought to the life of people. However, with the development of mobile communication technology and the expansion of mobile network scale, the propagation environment of mobile communication wireless signals is becoming more complex, and various electromagnetic signals in wireless channels interfere with the mobile network, thereby causing the overall degradation of wireless communication quality and affecting the service quality and user experience of the whole network.

Interference in a mobile communication network is a key factor influencing communication service quality, and can be divided into intra-network interference and extra-network interference according to the relationship between an interference source and an interfered network. The intra-network interference refers to interference caused by network element equipment with the same or similar parameter configuration to a signal, such as co-channel interference and adjacent channel interference, and the generated reason is that the limitation of wireless channel resources cannot meet the continuous increase of the network scale; the off-network interference refers to the interference situation caused between network element devices in different networks. Such as interference caused by the WLAN network to the LTE network, or interference signals generated by an off-network independent interfering device, such as a signal masker, etc. When these different network element devices happen to have identical or similar signal frequency points, spurious, blocking or intermodulation interference, etc. may be generated.

For the two different types of interference, different processing methods are available: for intra-network interference, configuration parameters of network element devices, such as transmission frequency, power, PCI parameters, and the like, can be optimally adjusted, and the related technologies are mature at present. For the off-network interference, the interference can be found in a way of drive test frequency sweep, and the like, however, because the position of the interference source is unknown, the rapid positioning of the position of the interference source is the most main task in the off-network interference analysis work. Only when the specific position of the interference source is determined, corresponding interference elimination processing measures can be taken to eliminate the influence of the interference on the network.

At present, two types of methods are mainly adopted for solving the problem of the positioning of the external interference in the mobile communication network: the first method is to adopt a test instrument to obtain various wireless measurement information about an interference source and a measurement receiving point, and to adopt various automatic positioning algorithms to estimate the position of the interference source according to the information. Available measurement information includes: angle of Arrival (AOA), Time of Arrival (TOA) and Time Difference of Arrival (TDOA), Received Signal Strength (RSS), channel state, phase amplitude, antenna array Received Signal, and the like. The adopted positioning method/algorithm comprises geometric positioning, fingerprint positioning, matrix completion, compressed sensing, factor graph, Bayesian model, deep neural network and the like; the second type is a manual interference positioning and troubleshooting method, in which network test maintenance personnel position the specific position of an interference source step by step in an interference occurrence area by adopting a manual test and successive approximation mode similar to wireless direction finding, but the mode is time-consuming and labor-consuming.

However, when the method for interference elimination and positioning is applied to a large-scale mobile communication network such as 4G, 5G, power wireless private network and the like, the following limitations exist: (1) under the practical application scenes of mobile communication network maintenance and interference positioning investigation, most of network maintenance testers are routine test instruments such as a frequency scanner and a road test terminal in daily use, and complex and expensive instruments such as a vector signal analyzer are rarely used. The wireless signal measurement information collected by testing mainly includes receiving frequency, signal strength, signal-to-carrier interference ratio/signal-to-noise ratio and the like, and accordingly interference checking and positioning are carried out. However, the measurement data (such as the CSI, the antenna array signals, and the antenna array signals) required by some of the above automatic interference positioning algorithms cannot be obtained conveniently and at low cost, so that these methods cannot be applied to practice conveniently. (2) In a complex urban environment, buildings and terrain features cause non-line-of-sight (NLOS) propagation of wireless signals, such as reflection and diffraction of the signals, and the linear distance between a signal source and a signal measuring point and measurement indexes such as signal intensity, arrival time, signal-to-noise ratio and the like of the wireless signals received by the measuring point show a severe nonlinear relation. The automatic interference positioning algorithms such as the geometric positioning, the matrix completion, the compressive sensing and the like bring large positioning errors.

Therefore, in the face of a complex wireless propagation environment in a dense urban area, the non-line-of-sight propagation characteristic of a wireless signal is considered, and the position of the wireless interference source in the mobile communication network is quickly and accurately located by utilizing wireless measurement information such as received signal strength provided by conventional wireless measurement equipment, so that the manpower and material resource investment is reduced, and the working efficiency of network interference investigation is improved.

References (e.g. patents/papers/standards):

[1]Wang Y,Wang S,Deng Y.A modified efficiency centrality to identify influential nodes in weighted networks[J].Pramana,2019,92(4):68.

[2]Beni H A,Bouyer A.TI-SC:top-k influential nodes selection based on community detection and scoring criteria in social networks[J].Journal of Ambient Intelligence and Humanized Computing,2020:1-20.

[3]Zareie A,Sheikhahmadi A,Jalili M.Influential node ranking in social networks based on neighborhood diversity[J].Future Generation Computer Systems,2019,94:120-129.

[4]Zareie A,Sheikhahmadi A,Khamforoosh K.Influence maximization in social networks based on TOPSIS[J].Expert Systems with Applications,2018,108:96-107.

[5]Miao W,Ye W,Yin J,et al.Coverage Analysis in TD-LTE Wireless Private Networks for Power Systems:A 3D Ray-Tracing Approach[C]//2018IEEE International Conference on Big Data and Smart Computing(BigComp).IEEE Computer Society,2018.

[6]Wang Y,Cong G,Song G,et al.Community-based greedy algorithm for mining top-kinfluential nodes in mobile social networks[C]//Proceedings of the 16th ACM SIGKDD international conference on Knowledge discovery and data mining.2010:1039-1048.

disclosure of Invention

The invention provides a mobile communication network interference positioning method, which aims at a complex wireless transmission environment of a dense urban area, considers the influence of wireless signal non-line-of-sight transmission (reflection, diffraction and diffraction) on interference positioning precision and the mutual influence between adjacent positions in space caused by landforms such as slopes and mountains and surface features such as buildings in the environment aiming at a mobile communication network which is deployed and operated; the propagation path of the interference signal is tracked and analyzed by adopting the reverse ray, and the network interference source is automatically positioned based on the network propagation analysis technology, so that the ideal interference positioning precision is achieved, and meanwhile, the manpower and material resource investment of drive test and frequency sweeping is reduced.

The method for positioning the mobile communication network interference comprises the following specific steps:

step one, selecting part of branch measuring points from a sweep frequency road measuring point set RTSet ═ sp to form a reverse ray tracking starting point set

The selected drive test points should satisfy: the signal intensity sp.rect measured at the road test point can not be lower than-85 dbm; the number of buildings in a 15 m-15 m rectangular range taking the road measuring point as the center is limited to be not more than 3 by the barriers near the road measuring point; the distribution of the road measuring points should be dispersed as much as possible, and the distance between every two road measuring points is set to be larger than 60 m.

Step two, tracking each point sp in the starting point set SPSet from the backward ray_iStarting from, performing back ray tracing to simulate the simulation of sp from the interference source to each starting point_iSignal propagation track tk of_iThereby generating a set of inverse ray propagation trajectories TrajSet ═ tk_i}。

And step three, screening a plurality of tracks similar to the propagation track in the track set TrajSet, only reserving one representative propagation track, and removing other redundant tracks.

When the two tracks have the following characteristics, the two tracks are similar tracks; the method specifically comprises the following steps:

1) have the same start and end grids; 2) the same number of reflection times; 3) the collision point around the reflection each time is located in the same building.

For similar tracks, keeping the track with the minimum path loss as a representative main path propagation track;

step four, dividing the interference source search area into areas with the size of C_i×C_i×C_hFor each screened strip of the three-dimensional gridDetermining a three-dimensional grid through which the three-dimensional grid passes and recording a track connection set of adjacent grids;

the method specifically comprises the following steps: aiming at each track, determining each three-dimensional grid through which the track passes one by one, and if a certain three-dimensional grid passes through the track, calling the grid to cover the track; otherwise, the grid is said not to cover the trace; counting track information covered by each grid and establishing grid index data structure Info_gtAnd storing the track information covered by each grid.

Meanwhile, if one track continuously passes through the grid p and the grid q, the track is stored to the connecting array tgcnt_pqThe array records the total number of tracks and the type distribution of the tracks of two adjacent grids.

All the connection arrays constitute a trace connection data set TgCon ═ { tgcnt ═ tgcnt [ ]_pq}

Taking each three-dimensional grid as a node, wherein the position of the node is the central position of the three-dimensional grid, and forming a side between the nodes corresponding to the two grids through which the track passes continuously; the nodes and edges make up the network model of the entire interferer search area.

Each node has the same attributes as the stereo grid overlay ray information. Array tgcnt_pqRegarding the side information connecting the node p and the node q, the side information comprises the ray number and the ray type distribution contained in the side; the number of rays contained by an edge is equal to the number of tracks connecting nodes at two ends of the edge, and the type of the ray is equal to the type of the track containing the ray.

Sixthly, indexing data Info according to the grids_gtCalculating each attribute value related to each node in the network model;

the attribute values include the number of reachable starting points, the number of covered rays, the integrated interference signal difference value, and the like.

The reachable starting points are: if the grids corresponding to a certain node a and a certain back-tracking starting point b cover the same track, the back-tracking starting point b is called as the reachable starting point of the node a.

The number of rays covered refers to: the number of tracks passing through the node;

node V_iCombined interference ofSignal difference V_iCISDif, the calculation formula is as follows:

wherein TPSet is node V_iA set of reachable starting points of (a); n is the pairwise combination number of the TPset set, and the value is

traj_iIs from a starting point sp_iDeparture, arrival node V_iTrace of minimum propagation loss; traj_jIs from a starting point sp_jDeparture, arrival node V_iThe trace with the smallest propagation loss.

And step seven, forming an initial node set OriVSet by all nodes in the network model, filtering out nodes which cannot be an interference source according to a filtering rule, and generating a candidate node set CanVSet by the rest nodes.

The filtering rules include:

1) filtering out nodes around the back tracking starting point and within a distance of less than or equal to 60m from the back tracking starting point;

2) filtering out nodes with the reachable starting points less than or equal to 50% of the total number of the back tracking starting points;

3) ratio of strong signal path_SrayFiltering nodes less than or equal to 25 percent;

the strong signal path occupation ratio calculation formula of the node is as follows:

Ratio_Sray＝ray_d/ray_sum

wherein ray_dNumber of direct paths covered by nodes, ray_sumIs the total number of paths covered by the node.

Step eight, calculating the influence of each node of the candidate node set CanVSet and the propagation probability of all edges corresponding to each node;

node V_iThe formula for calculating the influence is as follows:

f₁(V_i) Represents a node V_iNumber of reachable starting points of f₂(V_i) Represents a node V_iNumber of rays covered, f₃(V_i) Represents a node V_iIntegrated interference signal difference, w_kThe weight corresponding to each attribute.

Via node V_iEdge W of_ijThe propagation probability of (c) is:

W_ij·numrepresenting the total number of rays, ConRay, contained by the edge_maxRepresenting the maximum value of the number of rays contained in all edges in the network model; beta is a parameter for adjusting the network weight.

And step nine, aiming at each node in the candidate node set CanVSet, carrying out influence propagation in the network model by taking the node as a seed node, and selecting the node with the maximum propagation influence as an interference source positioning position.

In particular, for the current seed node V_iBy a propagation probability W_ij·probInfluencing the corresponding neighbor node V_jNeighbor node V_jSubsequently, the propagation probability W is also used_jp·probAffecting its neighbor node V_pAnd the rest is done until the set propagation depth is reached.

Counting the seed node V in the transmission process_iAll the affected nodes calculate the sum of the influence of all the affected nodes as the propagation influence V of the seed node_i·tinf。

Similarly, each seed node in the candidate node set CanVSet is subjected to influence propagation according to the process to obtain respective propagation influence.

And finally, selecting the seed node corresponding to the maximum value from all the propagation influence forces as the positioning position of the interference source.

The invention has the advantages that:

1) the mobile communication network interference positioning method analyzes the positioning of the unknown interference source outside the network based on the network propagation influence, and reduces the investment of manpower and material resources while ensuring that the ideal positioning precision is achieved.

2) The invention also discloses a mobile communication network interference positioning method, which considers the relevant indexes of a plurality of signal transmission and simultaneously considers the relation between adjacent positions, thereby further improving the positioning accuracy.

Drawings

Fig. 1 is a flow chart of a mobile communication network interference positioning method according to the present invention;

FIG. 2 is a diagram of interference source location based on network propagation impact analysis according to the present invention;

FIG. 3 is a diagram illustrating the definition of an interference source detection zone according to the present invention;

FIG. 4 is a flow chart of the present invention illustrating back ray tracing;

FIG. 5 is a diagram of a network impact modeling process of the present invention;

FIG. 6 is a flow chart of the node filtering algorithm of the present invention.

Detailed Description

The invention is described in further detail below with reference to the figures and examples in order to facilitate the understanding and implementation of the invention by those of ordinary skill in the art.

The invention is oriented to dense urban areas, and aims at TD-LTE networks which are put into use, irregular terrain and ground objects such as slopes and buildings exist in an interference search area, so that interference signals collide with the terrain and ground objects in the propagation process, and various conditions such as direct projection, reflection or diffraction exist, and the same measuring point can receive multiple paths from an interference source, namely, the propagation of the signals has a multipath phenomenon. Considering the influence of the complexity of the environment on signal propagation and considering the actual signal propagation, the invention provides a mobile communication network interference positioning method, starting from strong signal path measuring points with wide distribution and less shielding, simulating the propagation track of an interference signal, modeling grids and signal tracks in an interference search area by analyzing the path distribution and propagation loss, providing a standard taking a propagation influence index as interference source positioning, designing a network influence propagation model for influence propagation, and finally realizing the interference source positioning by judging the propagation influence of nodes.

As shown in fig. 1, the method for locating the interference of the mobile communication network specifically includes the following steps:

The simulation interference signal required by the positioning of the interference source is generated by a reverse ray tracking starting point, and the reverse ray tracking starting point is selected from the driving test points; in order to ensure that the generated simulated interference signal is more accurate and ensure the accuracy of the subsequent interference positioning stage, it is necessary to select some branch measurement points sp from the sweep frequency drive measurement point set RTSet { sp } according to the set conditions_iAnd the selected drive test points meet the following conditions: the signal intensity sp.rect measured at the road test point can not be lower than-85 dbm; the number of buildings in a 15 m-15 m rectangular range taking the road measuring point as the center is limited to be not more than 3 by the barriers near the road measuring point; the distribution of the road measuring points should be dispersed as much as possible, and the distance between every two road measuring points is set to be larger than 60 m.

The judgment rule that two tracks have similar propagation behaviors comprises the following steps:

And for each trace set with similar propagation behaviors, the trace with the minimum path loss is reserved as a main path, so that positioning interference caused by excessive similar path data is avoided.

Step four, dividing the interference source search area into areas with the size of C_i×C_i×C_hDetermining the three-dimensional grids passed by each screened track, and recording track connection sets of adjacent grids;

before a signal is transmitted reversely, an acceleration structure such as a uniform grid and a KD tree is adopted to model buildings, concave-convex terrains and the like in an interference source positioning area, and a space structure is recorded, so that a space environment during signal transmission is constructed; further, by analyzing the coverage condition of the TrajSet by the grid, grid index data Info is obtained_gt。

The method specifically comprises the following steps: aiming at each track, determining each three-dimensional grid through which the track passes one by one, and if a certain three-dimensional grid passes through the track, calling the grid to cover the track; otherwise, the grid is said not to cover the trace;

counting track information covered by each grid and establishing grid index data structure Info_gtAnd storing the track information covered by each grid.

And carrying out network modeling on the whole interference source searching area, wherein the network modeling comprises a plurality of nodes and edges connecting the nodes, and the corresponding relation between the nodes and the three-dimensional grid is as follows: for each three-dimensional grid, a unique node is corresponding to the three-dimensional grid in the network, the position of the node is the central position of the three-dimensional grid, and the node has an overlapping shot with the three-dimensional gridThe line information is of the same nature. If any track continuously passes through the two grids, a side is formed between the nodes corresponding to the two grids; array tgcnt_pqRegarding the side information connecting the node p and the node q, the side information comprises the ray number and the ray type distribution contained in the side; the number of rays contained by an edge is equal to the number of tracks connecting nodes at two ends of the edge, and the type of the ray is equal to the type of the track containing the ray.

node V_iOf the combined interference signal difference V_iCISDif, the calculation formula is as follows:

The filtering rules include:

Ratio_Sray＝ray_d/ray_sum

node V_iThe formula for calculating the influence is as follows:

Via node V_iEdge W of_ijThe propagation probability of (c) is:

Counting the seed node V in the transmission process_iAll the influencing neighbor nodes calculate the sum of the influences of all the neighbor nodes as the propagation influence V of the seed node_i·tinf。

Firstly, simulating an interference signal propagation path of an interference source by adopting a reverse ray tracing technology, and secondly, modeling the whole positioning area and the signal propagation path as a network through propagation path analysis; particularly, grid division is carried out on an interference source search area, grid-path index data are generated, grids correspond to nodes and edges in a network, and corresponding solving rules are formulated for various attribute values of the nodes and the edges; and then, carrying out redundancy filtering on the node and edge setting rules, and finally completing network modeling.

The specific process comprises the following stages:

the first stage is as follows: generating a set of inverse ray tracing starting points SPSet ═ sp_i}；

And a second stage: simulating a propagation path of an interference signal by a back ray tracking starting point;

and a third stage: and analyzing the propagation path, and combining the path data covered by the grid to complete network modeling.

The method mainly comprises the following steps:

first, Info for raster index data_gtAnd filtering to reduce the time for network modeling and improve the positioning accuracy of the interference source. By multiple paths similar to the propagation pathAnd screening is carried out, only one representative propagation path is reserved, and other redundant paths are removed.

And secondly, modeling the network node.

For node V in the network_iEstablishing the relation between the grid and the interference source positioning area grid, and indexing data Info according to the grid_gtThe path analysis of (2) calculates to obtain a node V_iAnd related item attribute values, thereby generating an initial candidate node set OriVSet.

And thirdly, modeling the network edge W.

For edge W in the network_ijEstablishing the connection between the signal propagation tracks and each signal propagation track of the connection nodes in the area, and indexing data Info according to the pair grids_gtBy path analysis of (3), calculating to obtain W_ijThe associated attribute values of the items.

And fourthly, filtering the initial candidate node set OriVSet, and reducing the scale of the node set to accelerate the subsequent influence propagation process and improve the positioning accuracy.

For OriVSet, a filtering rule related to node attributes is set, a plurality of nodes which are basically impossible to serve as interference sources are filtered, and a candidate node set CanVSet is generated.

Then, an influence propagation model is provided for the modeled network, the propagation condition of signals and the influence of adjacent nodes in an actual scene are comprehensively considered, a corresponding propagation rule and a propagation influence solving rule are formulated, and finally the positioning position of the interference source is determined according to the magnitude of the propagation influence of each node.

The method specifically comprises the following steps:

for each node V in the candidate node set CanVSet_iAnd carrying out influence propagation in the network by taking the node as a seed node, wherein the propagation probability is determined by a propagation probability attribute W.prob of the edge, and the attribute is calculated in a network modeling part. After a complete influence propagation process, the final propagation influence V of the seed node can be calculated_iTinf. And finally confirming the node with the largest propagation influence from all the candidate nodes as the interference source positioning position.

Example (b):

in the following, TD-LTE network is taken as an example, but the positioning method of the present invention is also applicable to other mobile cellular networks, such as FDD-LTE, WCDMA, and wireless private networks such as power wireless private network and trunking system.

First, an interference source search area is divided into a size C_i×C_i×C_hThe three-dimensional grid converts the problem of positioning the interference source into the problem of positioning the grid where the interference source is located, after network modeling is completed, the one-to-one correspondence relationship between the nodes and the grid is established, and all subsequent related interference positioning is completed based on the nodes. When the ray generated is traced by the backward ray tk_iWhen the element TrajSet passes through 1 grid, the grid is called to cover the ray tk_i。

As shown in FIG. 2, the selected back ray tracing origin is sp_i(i ═ 1,2,3,4,5,6), and the building is Bd_i(i 1,2,3,4) with a dashed line from the back ray tracing start point sp_iThe emitted reverse tracking rays generated by simulating the interference signal propagation form an interference signal propagation path, the five-pointed star Insp is an actual interference source point, and the square CanV is_i(i ═ 1,2,3) are interference source candidate node positions obtained after screening, and constitute an interference source candidate position set CanVSet ═ CanV₁,CanV₂,CanV₃}。

The 3 interferer candidate positions CanV in fig. 2_iThe rays are all denser, but the signal strength and path loss information of the rays covered by the 3 nodes are different. And positioning a target interference source node targV from a candidate node set CanVSet by analyzing ray information covered by the nodes and utilizing an influence propagation model, wherein the central point of the targV is used as the position of the interference source. Due to CanV₂Has the greatest propagation impact among the three candidate nodes, so the positioning result node is CanV₂And the center point is closer to the Insp, so that the positioning result is ideal.

The positioning of the off-network interference source based on the network propagation influence analysis comprises the following 7 steps.

Step 1, defining an interference source search range and generating a three-dimensional wireless signal propagation environment

As shown in fig. 3, according to the mobile communication network coverage area a_oDetermining an area A by information provided by an internal network management system, such as base station interference statistical information (such as bottom noise and PRB interference), a drive test sweep frequency analysis result obtained in advance, user complaints and the like_oThe position of interference occurs in the interference area, and then the detection range of an interference source is defined;

the method specifically comprises the following steps: firstly, evaluating the interfered degree of the interfered cell/base station according to the interference statistical information of the base station, and screening the interfered degree exceeding the interference threshold value thld_eintThe cell/base station, which is placed in the interfered point set inp set.

Then, the drive test sweep frequency data is analyzed, and the interference signal intensity exceeding the interference threshold value thld is screened out_eintThe swept path measurement point is placed in the interfered point set InPSet.

And then, according to the user complaint data, analyzing the position of the complaint point and the interference severity degree, thereby determining the position point which is seriously interfered in the network and putting the position point into the interfered point set InPSet.

Finally, constructing a rectangular bounding box A according to the positions of all interference points in the InPSet₁And expanding the set distance sd to the periphery to obtain a rectangle A, namely the interference source detection area.

Since the ray emitted from the emission starting point collides with the rugged topography in the environment during the propagation process, in order to simulate the collision between the actual ray and the real environment, the obstacle existing in the three-dimensional environment needs to be represented and stored, that is, the three-dimensional environment is generated by simulation. Taking the interference source detection area a as an example, recording the space environment of the whole area a by artificially constructing a uniform grid acceleration structure.

The specific construction and representation process is as follows:

the first step, equally dividing the three-dimensional space of the area A into C_i×C_i×C_hEach cuboid being referred to as a grid Tg_i(ii) a Equally dividing the ground of area A into sizes C_i×C_iA square area of (a); in order to indicate the position of the grid by the grid number i, the number i and three dimensions need to be establishedThe relationship between the spatial coordinates. With the point O with the largest longitude and latitude in the area a as the origin of coordinates, O (0,0, 0). Respectively extending from the point O to the three directions of the length, the width and the height of the cubic area A, establishing an x-y-z three-dimensional coordinate system and establishing a grid Tg_iNumber i of (2) and its center point coordinate (Tg)_i·x,Tg_i·y,Tg_i·z) The conversion rules between:

α:i→(Tg_i·x,Tg_i·y,Tg_i·z),β:(Tg_i·x,Tg_i·y,Tg_i·z)→i

secondly, processing the model data containing the terrain height of the area A, vectorizing and triangularizing the model data to obtain a plurality of triangular data, and distributing a unique number tri to each triangular data_iThe convex terrain and the like can be represented by using the triangle data.

For a building, firstly, obtaining map data of an area A, and further obtaining building data; the buildings are then represented, for each of which there corresponds a unique number bd_iIf the buildings are uniformly regarded as irregular multi-face cylinders, the coordinates and the height of each vertex of the bottom face of each building are only needed to be stored.

And thirdly, generating grid acceleration information. In order to judge the collision condition with the obstacle more quickly in the subsequent ray propagation stage and reduce the complexity of collision detection, the building and terrain information need to be uniformly represented.

For each grid Tg_iIf its ground projection is in contact with the building bd_jIntersect or are surrounded by the bottom surface, then bd_jAddition of grid Tg_iAcceleration information structure Tg of_iacSet. Similarly for convex topography, if grid Tg_iGround projection quilt tri_kEnclosing or intersecting with, then will tri_kAddition of grid Tg_iAcceleration information structure Tg of_iacSet.

Through the steps, the simulation generation of the three-dimensional wireless signal propagation environment is completed, and meanwhile, the environment information is recorded and expressed so as to accelerate the subsequent ray propagation process.

Step 2, analyzing the terrain and ground feature environment information to generate an initial measuring point

After the three-dimensional wireless signal propagation environment simulation generation is carried out in the last step, initial measurement points are selected according to the established area A environment information, and the drive test is planned to generate relevant data of the measurement points.

The specific process is as follows:

first, when selecting a measurement point, each point in the area a should be within the effective coverage area of the measurement device in consideration of the limitation of the signal measurement device due to the unknown specific location of the interference source. If the measuring equipment has an effective measuring radius r, at least two sp(s) with the sp being less than or equal to r need to be satisfied for any point p in the area A so as to meet the precision requirement of subsequent positioning; through this step, an initial measurement point set RTSet { sp } is generated.

And secondly, planning a measuring track. And planning a traveling track of the measuring equipment according to the position of the measuring point acquired in the first step. Firstly, the measurement track is ensured to cover all measurement points in the RTSet, and secondly, the total length of the measurement track is reduced as much as possible so as to reduce the consumption of manpower and material resources.

And thirdly, recording relevant information of the measuring points. Performing field measurement according to the measurement track, and recording the measurement result at each measurement point sp_iMeasurement data obtained from position, including current time, signal strength sp_iRecitr and frequency band, and the like, and simultaneously acquires the longitude and latitude at the measuring point and records the data into the attribute field corresponding to the measuring point.

Step 3, selecting a part of signal transmitting starting points and transmitting initial rays;

and 3.1, screening the initial measurement point set RTSet according to a certain rule, and selecting and generating a final signal transmission starting point set SPSet.

The core principle is to more accurately simulate the strong interference signal emitted by an actual interference source by using a transmitting starting point. According to this principle, firstly, the signal strength measured at the transmission starting point cannot be too small, otherwise, it indicates that the propagation loss of the signal reaching the point via the interference source is too large, and the number of the received strong interference signals is substantially zero. Secondly, the number of obstacles near the emission starting point cannot be too large, otherwise, too many collisions can be generated in the propagation process, so that the number of weak signals is too large, and the positioning accuracy is affected. Finally, in order to reduce redundant signal tracks and improve subsequent positioning efficiency, the emission starting points should be dispersed as much as possible.

The specific measurement point screening process is as follows:

the first step, initializing set SPSet ═ phi, and setting signal strength threshold sigstr_minFor each sp_iBelongs to RTset, if sp is satisfied_i·recstr≥sigstr_minThen, update SPSet: sp is to be_iAdded to the set SPSet.

Second, setting the scanning radius r_scanAt each sp_iSimulating the generation of the length r from the start of the RTSet_scanOf (2) is performed. If the rays with a large angle range collide with buildings or terrains, sp is considered_iSurrounded by obstacles, otherwise update SPSet: sp is to be_iAdded to the set SPSet.

Third, setting the shortest distance dis_minAnd minimum area_min. Updating SPSet: the emission points in the SPSet were screened so that all points remaining met: 1) for arbitrary two points sp_i,sp_jE is the SPSet; i ≠ j, its distance d (sp)_i,sp_j)≥dis_min(ii) a 2) For arbitrary three points sp_i,sp_j,sp_kE is the SPSet; i ≠ j ≠ k, and its triangular area S (sp) formed_i,sp_j,sp_k)≥area_min。

And 3.2, simulating and generating a possible interference source signal track aiming at the signal transmission starting point set SPSet.

Taking into account the emission start point sp_iIt may receive both direct and indirect signals from an interference source, and thus split into two parts when emitting primary rays: for the direct signal part, from sp_iEmitting towards the periphery and the top boundary of the positioning area A; for the non-direct signal part, from sp_iTo the building surface in area a.

Using the structure Pseq { < pinf > } PairRecording the end points of all sections of rays in a complete propagation path; pinf represents a path from a starting point sp_iTo an ordered set of reflection, diffraction points, and end points. The specific attributes include the coordinates and type of each point.

Recording information of each section of rays in a complete propagation path by using a structure PathSet { < ray > }, wherein < ray > represents an ordered sequence of each section of rays when propagating from a starting point to an end point in the path; specific attributes include the type, level, length, attenuation level, etc. of the ray.

For slave sp_iPrimary ray emitted, adding sp_iAnd (4) information is less than pinf, including the position information and the signal strength of the information, so as to carry out the subsequent ray tracing process.

Step 4, analyzing the ray propagation track to generate a signal propagation path

And performing a subsequent ray tracing process on the generated primary ray. As shown in fig. 4, the specific steps are as follows:

first step, for the current ray_iThe propagation path is analyzed, and if the ray directly exits from the boundary portion of the area a, the ray tracing process is terminated. Calculating the point of intersection fp, ray, at the time of injection_iAnd adding to < pinf > the endpoint coordinate information of (1). At this time, a complete sp from a certain emission starting point is saved in the Pseq set_kStarting to a signal track with the termination of the point fp at the intersection point, and going to the fifth step; if the ray does not exit from the boundary, go to the second step.

Second step, ray for current ray_iJudging whether the attenuation is excessive or not; calculating the current ray_iFor a preset maximum number of collisions, cracnt_maxIf N > cracnt_maxAnd (4) the signal strength is over-lost, the subsequent propagation process is not tracked, and the step five is carried out, otherwise, the step three is carried out.

Third, for the current ray_iAnd the Tg of the grid in which it is currently located_jTo judge ray_iWhether or not it is in contact with Tg_jThe ground features in (1) collide.

In particular, the grid is traversedTg_jAcceleration information structure Tg of_jAll buildings bd recorded in acSet_iAnd the topography tri_iData, and ray_iAnd performing collision detection. If ray_iAnd (4) intersecting with the building or the surface of the raised terrain, recording the position information of the crop point crop of the intersection point, and turning to the fourth step. Otherwise, the ray tracing is continued to reach the next grid, and the second step is carried out.

Fourthly, obtaining ray according to the position information of the crop of the intersection point_iWhether it is reflected or diffracted. If the crop is on the surface of the polygon, then ray is determined_iA reflection occurs and crop related reflection information is recorded and added to < pinf >. Generating the next-stage reflection ray_i+1，ray_i+1Is transmitted in the direction of travel by ray_iIs calculated from the propagation direction of the light beam and the collision surface data, and is used to calculate the reflection ray_i+1And continuing to track and returning to the second step. If the crop is positioned at the edge of the polygon, then the ray is judged_iDiffraction occurs and crop related diffraction information is recorded and added to < pinf >. Generating the next level of ray_i+1The same principle as that of the reflected ray is used to obtain ray_i+1And the direction of propagation of the contra-wound ray_i+1And continuing to track and returning to the second step.

Fifthly, according to the data of < pinf > in the Pseq, constructing sp from the emission point one by one_iEach section of ray in the emitted signal track₀，ray₁，…，ray_k. During construction, the starting and ending point positions of all sections of rays, the number of stages and the reflection/diffraction loss coefficient need to be recorded. All the ray data together form a sp_iThe complete signal path issued < ray >.

For all transmission start points sp in the set of transmission start points SPSet_iAll slave sp are generated by < pinf > data in Pseq_iThe transmitted signal propagates the path and finally constitutes the set of all paths PathSet { < ray > }.

Construction of trellis-Path index Info with all propagation Signal Path data < ray >_gt: the calculation process is to use the propagation direction of each ray in < ray > to access the passing grid Tg one by one_iThereby obtaining each grid Tg_iInformation such as the source point of the covered ray signal, the propagation loss and the like is recorded in the index structure Info_gtIn (1).

Step 5, analyzing the distribution of the propagation paths to complete the modeling of the influence network

The core of the invention is that the network modeling is carried out by the grids divided in the whole interference source searching area and the generated path track, and a certain influence is given to the nodes in the network. And then carrying out influence propagation on the network by using an influence propagation model, thereby calculating the sum of the influences caused by propagation of each node, and selecting the node with the maximum propagation influence from the sum as the node where the positioned interference source is located.

The grid division is completed in the previous step, and the propagation of the interference signal is simulated by using a reverse ray tracing model, so that a plurality of signal propagation paths are generated; further analyzing the propagation path and establishing a raster-path index Info_gt。

Step 5.1 the present invention regards the grid and path trajectory in the interference source search area as a network G ═ V, W, where node V corresponds to grid Tg and edge W corresponds to a path tk cluster connecting two nodes. In the network modeling including subsequent influence propagation steps, the nodes and grids are equivalent.

As shown in fig. 5 for grid Tg_i(i 1, 2.., 9), which are in one-to-one correspondence with nodes V in the network_i(i 1, 2.., 9) while the correlation properties of the signal propagation of the grid itself are converted into influence properties of the network nodes. For ray trajectory tk_iIf it has traveled successively through two adjacent grids Tg_iAnd Tg_jThen form a node V in the network_iAnd V_jA side W therebetween_i,jEach edge is constructed only once, and the number of tracks corresponding to one edge is converted into a weight corresponding to the edge.

The specific correspondence is as follows:

search for each grid Tg in region A for an interferer_iWhich uniquely corresponds to a node V in a network G_i. For any node V_iWhich isHaving a certain influence V in the network G_iW, the influence is a quantized value of path information covered by a grid corresponding to the node itself, and intuitively, the larger the influence, the more core the node occupies in the network. In particular by node V_iNumber of reachable starting points V_iAvaNum, number of covered strong signal paths V_iStrajNum and the integrated interference signal difference V_iCISDIF.

The comprehensive interference signal difference value of the grid refers to that the grid receives back tracking rays sent from a plurality of different path measuring points, the sum of grid transmitting power difference values obtained by back derivation of the rays sent from two paths of measuring points is counted, and the sum is equal to the comprehensive interference signal difference value of the node due to the correspondence between the grid and the node. The pseudo code for calculating the composite interferer difference for each node in the CanVSet based on the composite interferer difference definition may be as follows, and may pass V after the calculation is complete_iPloss Access node V_iThe integrated interference signal difference value:

the pseudo code process for calculating the integrated interference difference values of all nodes of the CanVSet is as follows:

inputting: raster-path index Info_gtAnd a set of interferer candidate locations CanVSet;

and (3) outputting: a comprehensive interference difference set CISdiffet of the nodes;

firstly, initializing a CISdiffet; for each node V in the set CanVSet_iCounting node V_iA set tmpSet of reachable starting points of (c); initializing diff_ploss:＝0；

For any two different starting points sp in the set tmpSet_iAnd sp_jCalculating the starting points sp respectively_iAnd sp_jTo node V_iSignal propagation path traj with minimum propagation loss_iAnd traj_j(ii) a And update diff_ploss+＝Abs(str(V_i|traj_i)-str(V_i|traj_j) ); until the CISDifset.Add (V) is obtained_i,diff_ploss) And (6) outputting.

For any two nodes V in the network G_iAnd V_jIf they are pairedThe corresponding grids are spatially adjacent and cover at least one identical path, then the node V is called_iAnd V_jAdjacent to and having an edge w_i,jConnection V_iAnd V_j. For the edge w_i,jHaving an attribute w_i,jNum and w_i,jProb, wherein w_i,jNum denotes a connection node V_iAnd V_jNumber of paths of, w_i,jProb represents the node V in the subsequent impact propagation phase_iAnd V_jThe probability of propagation in between. Although the actually generated signal propagation path has directivity, the network G is undirected, i.e., w, because neighboring nodes can propagate each other when the influence propagates_i,jAnd w_j,iAre equivalent.

And 5.2, considering that the finally generated signal propagation paths are excessive, paths with similar propagation behaviors need to be screened, and a representative main path is finally reserved. The specific main path screening process is as follows:

first, initialize temporary main path index TmepInfo_gtFor null, the ground portion of the interferer search region is divided into R_l*R_wThe grid-path index Info_gtAnd the end points of the middle paths are projected on the ground, so that the two paths with the projection points located on the same rectangular block have the same end point.

Second step, for Info_gtEach path tk in_i. And if all paths have been traversed, turning to the fifth step. If at TmepInfo_gtTo find a path tk having the same starting and ending points as it had_jEntering the third step, otherwise updating TmepInfo_gtTo be tk_iAdded and transferred to the second step.

Third step, if tk_iAnd tk_jIf the number of the winding reflection times is the same, the fourth step is carried out, otherwise, the TmepInfo is updated_gtTo be tk_iAdded and transferred to the second step.

Step four, if tk_iAnd tk_jIf the sequence of the building in collision is not completely consistent, the TmepInfo is updated_gtTo tk_iAdded and transferred to the second step. If not, then,if tk_jPath loss ratio tk of_iSmall, go to the second step, if tk_iIs smaller, then tk is used_iPath coverage tk_jThe data of (1). And turning to the second step.

Fifth step, update Info_gt:＝TmepInfo_gt。

And 5.3, after the paths are screened, removing part of redundant paths, reducing errors caused by the redundant paths to a subsequent network modeling stage, and saving modeling time.

The specific network modeling process is as follows:

first, index Info with raster-Path_gtCalculating each node V of candidate node set CanV_iNumber of reachable starting points V of corresponding grid_iAvaNum, number of covered strong signal paths V_iStrajNum and the integrated interference signal difference V_i·CISDif。

Secondly, respectively counting the maximum value AvaNum of the reachable starting points covered by all grids_maxAnd median AvaNum_midMaximum StrajNum of number of covered strong signal paths_maxAnd median StrajNum_midThe minimum value of the interference signal difference is synthesized with the CISdiff_minAnd a median CISDif_mid。

Thirdly, obtaining an influence node V by using the calculation results of the first two steps_iWeight V_iThe calculation formula of the respective influence factors of w:

wherein f is₁(V_i) Number of reachable starts for influencing factorCalculation function, f₂(V_i) For the calculation function of the number of strong signal path coverages of the impact factor, f₃(V_i) The calculated function of the interference signal difference is integrated for the impact factor.

Fourthly, according to the importance degrees of the three influence factors, respectively setting corresponding weight coefficients w for the three influence factors_kTo obtain the final node V_iWeight V_iThe formula for w is as follows:

fifth step, for node V_iAnd node V_jEdge W between_ij. Traversal raster-path index Info_gtData, each having a signal propagation path tk_iPassing through V continuously_iAnd V_jThen update W_ijNum plus one; and obtaining the number of ray connections represented by all edges after the traversal is finished.

Sixthly, solving the maximum ConRay of the number of the connecting rays represented by all edges in the network_maxFor all edges, with W_ijFor example, it represents a propagation probability W_ijProb is calculated as follows:

the parameter beta is used for adjusting the propagation probability trend of all edges in the network, and if the probability is larger than 1, the propagation probability is set to be 1.

Step 6, determining the position of the interference source by adopting a network influence propagation model

And aiming at the candidate node set CanVSet, spreading the influence of each node in the CanVSet by using an influence spreading model, and selecting the node with the maximum spreading influence as a positioning position of the interference source.

The method specifically comprises the following steps:

step 6.1, obtaining an initial node set OriVSet through a node modeling part, considering that the OriVSet is overlarge in scale, in order to reduce time consumption in a subsequent influence propagation stage, filtering and screening the OriVSet, removing nodes which are basically not used as interference source positions, and generating a candidate node set CanVSet; FIG. 6 shows a filtering process for an initial set of nodes.

Specifically, an iterative algorithm is set for an initial candidate node set OriVSet, and finally a candidate set size meeting an expected range is generated, wherein the filtering algorithm flow is as follows:

firstly, initializing each parameter of an iterative algorithm: including the maximum iteration number Cnt, the reachable starting point number filtering threshold AvaNum_minThe covered strong signal path fraction filter threshold Sprate_minCurrent iteration number CurCnt, target candidate node set size range [ Scale_min,Scale_max]。

Second, initialize CanVSet ═ OriVSet, and update CanVSet: nodes around the start of reverse transmission in the CanVSet are removed.

And step three, updating the CanVSet: if there is node V in the set_i·AvaNum＜AvaNum_minThen V will be_iRemove from CanVSet and update V_i·w＝0。

Step four, updating the CanVSet: if there is node V in the set_iThe ratio of covered strong signal paths is less than the SPrate_minThen V will be_iRemove from CanVSet and update V_i·w＝0。

And fifthly, updating the current iteration frequency CurCnt. If the iteration number exceeds the maximum iteration number Cnt, or the number of nodes in the candidate node set CanVSet is in the range [ Scale_min,Scale_max]And if so, ending the iteration, otherwise, turning to the sixth step.

Sixthly, if the situation is that CanVSet is less than Scale_minIf the number of filtering nodes is too large, the filtering condition needs to be properly relaxed to reduce AvaNum_minAnd SPrate_minThe value of (c). If | CanVSet | > Scale at this time_minIf the number of filtering nodes is too small, the AvaNum needs to be increased appropriately_minAnd SPrate_minThe value of (c). And returns to the second step.

Step 6.2 byThe filtering process can remove a large number of candidate nodes, and for the removed node V_iThe influence is updated to 0, and the subsequent influence propagation stage does not have the propagation capacity, so that the propagation time is greatly shortened.

In the propagation model, each node may be in an active state or an inactive state at a certain time, the inactive state node is activated and then transits to the active state, and the active state node cannot be activated again. In the propagation process, an active state node set S at the current time t is mainly maintained_tAnd if the active node set is not changed from the time t to the time t +1, ending the transmission process in the current round.

For node V_iHaving a propagation influence V_iTinf, the specific propagation procedure is as follows:

the first step is to initialize a node set TempVSet ═ CanVSet to be transmitted.

And secondly, if the TempVSet is empty, the propagation process is ended. Otherwise, optionally V_iE.g. TempVSet as seed node of the current round of propagation, and V_iDeleted from TempVSet. Initialization S₀:＝{V_i}，

t:＝0，V_i·tinf:＝V_i·w。

Third, set S is initialized_t:＝S_t-1For all V_j∈S_t-1\S_t-2For each of which is not in the set S_t-1Neighbor of (1)

V_jWith a probability W_jkProb is activated if V_kIs successfully activated, V is activated_kIs added to S_tIn and update V_i·tinf:＝V_k·w+V_i·tinf。

Fourth, judging S_tWhether or not to react with S_t-1And if the two are equal, the transmission process of the round is finished, and the step is switched to the second step. And if not, updating t: (t + 1) and turning to the third step.

After the influence propagation process, the propagation influence of all nodes in the candidate node set CanVSet is obtained, and the propagation influence V is selected from the propagation influence_tarNode V with the largest tinf_tarAs the location of the interference source.

In the process of positioning the off-network interference source based on the network propagation influence analysis, the network modeling is carried out aiming at the grid and the signal track, and the network is generated or different influence propagation models are utilized according to other indexes or parameters based on the influence propagation model and the like, thereby still belonging to the protection range of the application.

Claims

1. A mobile communication network interference positioning method is characterized by comprising the following specific steps:

step one, selecting partial branch measuring points from a sweep frequency road measuring point set to form a reverse ray tracking starting point set;

step two, tracking each point sp in the starting point set from the backward ray_iStarting from, performing back ray tracing to simulate the simulation of sp from the interference source to each starting point_iSignal propagation track tk of_iThereby generating a set of inverse ray propagation trajectories TrajSet ═ tk_i}；

Step three, screening a plurality of tracks similar to the propagation track in the track set TrajSet, only keeping the track with the minimum path loss, and removing other redundant tracks;

taking each three-dimensional grid as a node, wherein the position of the node is the central position of the three-dimensional grid, and forming a side between the nodes corresponding to the two grids through which the track passes continuously; the nodes and the edges form a network model of the whole interference source searching area;

the attribute values comprise the number of reachable starting points, the number of covered rays and the difference value of the comprehensive interference signal;

step seven, forming an initial node set OriVSet by all nodes in the network model, filtering out nodes which cannot be an interference source according to a filtering rule, and generating a candidate node set CanVSet by the rest nodes;

node V_iThe formula for calculating the influence is as follows:

f₁(V_i) Represents a node V_iNumber of reachable starting points of f₂(V_i) Represents a node V_iNumber of rays covered, f₃(V_i) Represents a node V_iIntegrated interference signal difference, w_kWeights corresponding to the attributes;

via node V_iEdge W of_ijThe propagation probability of (c) is:

W_ij·numrepresenting the total number of rays, ConRay, contained by the edge_maxRepresenting the maximum value of the number of rays contained in all edges in the network model; beta is a parameter for adjusting the network weight;

2. The method as claimed in claim 1, wherein in the first step, the selected drive test point satisfies the following conditions: the signal intensity sp.rect measured at the road test point can not be lower than-85 dbm; the number of buildings in a 15 m-15 m rectangular range taking the road measuring point as the center is limited to be not more than 3 by the barriers near the road measuring point; the distribution of the road measuring points should be dispersed as much as possible, and the distance between every two road measuring points is set to be larger than 60 m.

3. The method as claimed in claim 1, wherein in the third step, the similar trajectories include the following features:

4. The method for locating interference in a mobile communication network according to claim 1, wherein the fourth step is specifically: aiming at each track, determining each three-dimensional grid through which the track passes one by one, and if a certain three-dimensional grid passes through the track, calling the grid to cover the track; otherwise, the grid is said not to cover the trace; counting track information covered by each grid and establishing grid index data structure Info_gtStoring the track information covered by each grid;

meanwhile, if one track continuously passes through the grid p and the grid q, the track is stored to the connecting array tgcnt_pqThe array records the total number of tracks of two adjacent grids and the type distribution of the tracks;

all the connection arrays constitute a trace connection data set TgCon ═ { tgcnt ═ tgcnt [ ]_pq}。

5. The method as claimed in claim 1, wherein in the fifth step, each node has the same attribute as the information of the coverage ray of the stereoscopic grid;

array tgcnt_pqRegarding the side information connecting the node p and the node q, the side information comprises the ray number and the ray type distribution contained in the side; the number of rays contained by an edge is equal to the number of tracks connecting nodes at two ends of the edge, and the type of the ray is equal to the type of the track containing the ray.

6. The method for locating interference in a mobile communication network according to claim 1, wherein in step six,

the reachable starting points are: if the grids corresponding to a certain node a and a certain back tracking starting point b cover the same track, the back tracking starting point b is called as the reachable starting point of the node a;

7. The method as claimed in claim 1, wherein in the seventh step, the filtering rule comprises:

Ratio_Sray＝ray_d/ray_sum

8. The method for locating interference in a mobile communication network according to claim 1, wherein the ninth step is specifically:

for the current seed node V_iBy a propagation probability W_ij·probInfluencing the corresponding neighbor node V_jNeighbor node V_jSubsequently, the propagation probability W is also used_jp·probAffecting its neighbor node V_pAnd so on until reaching the set propagation depth;

counting the seed node V in the transmission process_iAll the affected nodes calculate the sum of the influence of all the affected nodes as the propagation influence V of the seed node_i·tinf；

Similarly, each seed node in the candidate node set canVSet is subjected to influence propagation according to the process to obtain respective propagation influence;